Review of the Stratigraphy, Depositional Systems ...

73 downloads 58 Views 6MB Size Report
The Ellenburger Group of the West Texas Permian Basin is part of a Lower. Ordovician carbonate platform that covered large areas of the United States. During ...
DRAFT REVIEW OF THE LOWER ORDOVICIAN ELLENBURGER GROUP OF THE PERMIAN BASIN, WEST TEXAS Robert Loucks Bureau of Economic Geology Jackson School of Geosciences The University of Texas at Austin Austin, TX ABSTRACT The Ellenburger Group of the West Texas Permian Basin is part of a Lower Ordovician carbonate platform that covered large areas of the United States. During the Lower Ordovician, the Permian Basin area was located on the southwest edge of the Laurentia plate between 20 to 30 degrees latitude. The equator crossed northern Canada situating Texas in a tropical to subtropical latitude. Texas was a shallow-water shelf with deeper water conditions to the south where it bordered the Iapetus Ocean. Shallow-water carbonates were deposited on the shelf and deep-water shales and carbonates were deposited on the slope and in the basin. The interior of the shelf produced restricted environments, while the outer shelf produced open-marine conditions. The diagenesis of the Ellenburger Group is extremely complex and the processes that produced the diagenesis covered million of years. Three major diagenetic processes strongly affected the Ellenburger carbonates: (1) dolomitization, (2) karsting, and (3) tectonic fracturing. Pore networks in the Ellenburger are especially complex

1

because of the amount of brecciation and fracturing associated with karsting. The pore networks can consist of any combination of the following pore types depending on depth of burial: (1) matrix, (2) cavernous, (3) interclast, (4) crackle/mosaic breccia fractures, or (5) tectonic-related fractures. The Ellenburger Group is an ongoing important exploration target in West Texas. The carbonate depositional systems within the Ellenburger Group are relatively simple; however, the diagenetic overprint is extremely complex producing strong spatial heterogeneity within the reservoir systems. INTRODUCTION The Ellenburger Group of the Permian Basin is part of a Lower Ordovician carbonate platform that covered large areas of the United States (Figures 1, 2) (Ross, 1976; Kerans, 1988, 1990). It is well known for being one of the largest shallow-water carbonate platforms in the geological record (covering thousands of square miles and up to 500 miles wide in West Texas), being extensively karsted at the Sauk unconformity, and its widespread hydrocarbon production. Hydrocarbon production ranges from as shallow as 856 ft in the West Era field in Cooke County to as deep as 25,735 ft in the McComb field in Pecos County. A review of the Ellenburger Group will aid in understanding the sedimentology and diagenesis that has lead to this widespread producing unit. The first inclusive studies of the Ellenburger Group were completed by Cloud et al. (1945), Cloud and Barnes (1948, 1957) and Barnes et al. (1959). These studies covered many aspects of the group ranging from stratigraphy to diagenesis and chemistry. Much has been learned since then about carbonate sedimentology and diagenesis and these new concepts were integrated into later studies by Kerans (1988, 1989, and 1999). Kerans’ studies covered regional geologic setting, depositional systems, facies analysis, depositional history, diagenesis, and paleokarsting. Many other papers have described the local geology of fields (e.g., Loucks and Anderson, 1980, 1985; Combs et al., 2003), outcrop areas (e.g., Goldhammer et al., 1992; Lucia, 1995, 1996; Loucks et al., 2004), or have elaborated on paleokarsting (e.g., Lucia, 1971, 1995, 1996;

2

Loucks and Anderson, 1985; Kerans, 1988, 1989, 1990; Loucks and Handford, 1992; Candelaria and Reed (1992); Loucks 1999; Loucks et al., 2004). The major objectives of this paper are to review the: (1) regional geological setting and general stratigraphy, (2) depositional systems, facies analysis, and depositional history, (3) general regional diagenesis, (4) reservoir characteristics, and (5) petroleum system. Much of the data are from published literature, however, new insights can be derived from integrating these data. REGIONAL GEOLOGICAL SETTING At the global plate scale during the Lower Ordovician, the West Texas Permian Basin area was located on the southwest edge of the Laurentia plate between 20 to 30 degrees latitude (Figure 1) (Blakey, 2005a, 2005b). The equator crossed northern Canada (Figure 1) situating Texas in a tropical to subtropical latitude (Lindsay and Koskelin, 1993). Much of the United States was covered by a shallow sea. Most of Texas was a shallow-water shelf with deeper water conditions to the south where it bordered the Iapetus Ocean. The Texas Arch (Figures 2), a large land complex, existed in north Texas and New Mexico. Ross (1976) and Kerans (1990) pointed out that the main depositional settings within the Permian Basin for the Ellenburger Group were the deeper water slope and the shallower water carbonate platform. In Figure 2, Ross (1976) showed the broad Lower Ordovician carbonate platform with the interior being dolomite and the outer area being limestone. Seaward of the limestone he showed black shale. Kerans (1990) interpreted Ross’s map in terms of depositional settings (Figure 3). The dolomite being the restricted shelf interior and the limestone being an outer rim of more open-shelf deposits. Seaward of the platform was a deeper water slope system (shales), which Kerans (1990) stated is represented by the Marathon Limestone. The Ellenburger Group in the southern part of Texas, where the deeper water equivalent strata would have been, was strongly affected by the Ouachita Orogeny when the South American plate was thrusted against the North American plate (Figure 2). The basinal and slope facies strata were destroyed or

3

extensively structurally deformed. Ellenburger Group facies cannot be traced south of the slope setting in southwest Texas because of the Ouachita Orogeny. Kerans (1990) recognized that several peripheral structural features affected the deposition of the Ellenburger sediments in the West Texas New Mexico area (Figure 4); however most of the platform was relatively flat. Major structural features in the area that formed after Early Ordovician time include the Middle Ordovician Toboas Basin and the Pennsylvanian Central Basin Platform (Galley, 1958). Structural maps of the top Ellenburger Group (Figure 5) and top Precambrian intervals (Figure 6) show the Ellenburger Group at a structural low in the area of the Permian Basin. In the Midland Basin area, the top of Ellenburger carbonate is as deep as 11,000 ft, shallower over the Central Basin Platform, and as deep as 25,000 ft in the Delaware Basin. Isopach maps (Figure 4) by the Texas Water Development Board (1972), Wilson (1993), and Lindsay and Koskelin (1993) show thickening into the area of the Permian Basin. GENERAL STRATIGRAPHY The Ellenburger Group is equivalent to the El Paso Group in the Franklin Mountains, the Arbuckle Group in northeast Texas and Midcontinent, the Knox Group in the Eastern United States, and the Beekmantown Group of the northeast United States. In West Texas the Ellenburger Group overlies the Cambrian Bliss subarkosic sandstone (Loucks and Anderson, 1980). In the Llano area, Barnes et al. (1959) divided the Ellenburger Group from the bottom to the top into the Tanyard, Gorman, and Honeycut Formations. Kerans (1990) compared the Llano stratigraphic section to the subsurface stratigraphy of West Texas (Figure 7). A worldwide hiatus appeared at the end of Lower Ordovician deposition creating an extensive second-order unconformity (SaukTippecanoe Supersequence Boundary defined by Sloss (1963); Figure 7). This unconformity produced extensive karsting throughout the United States and is discussed later in this report. In West Texas, the upper Middle Ordovician Simpson Group was deposited above this unconformity (Figure 7).

4

A general second-order sequence stratigraphic framework was proposed by Kupecz (1992) for the Ellenburger Group in West Texas (Figure 8). The contact between the Precambrian and the Lower Ordovician intervals represents a lowstand of sea level of unknown duration. The Bliss Sandstone sediments are in part lowstand erosional deposits (Loucks and Anderson, 1985). The lower second-order transgressive systems tract includes the Bliss Sandstone and the lower Ellenburger alluvial fan to interbedded shallow-subtidal paracycles. The second-order highstand systems tracts include the upper interbedded paracycles of peritidal deposits. The next second-order lowstand produced the Sauk-Tippecanoe sequence boundary. The detailed sequence stratigraphy of the Lower Ordovician of West Texas (Figure 9) was worked out by Goldhammer et al. (1992), Goldhammer and Lehmann (1996), and Goldhammer (1996). Their work was in the Franklin Mountains in far West Texas. They compared their work to other areas including the Arbuckle Mountains in Oklahoma (Figure 9). They divided the general Lower Ordovician section, which they call the Sauk-C second-order supersequence, into nine third-order sequences (Figure 9) based upon higher order stacking patterns. Each third-order sequence had duration of 110 million years. Goldhammer et al. (1992) stated that the origin and control of thirdorder sequences in the Lower Ordovician remain problematic because this period of time lacks evidence for major glaciation. In the Franklin Mountains, Goldhammer et al. (1992), Goldhammer and Lehmann (1996), and Goldhammer (1996) only recognized the lower seven sequences (Figure 9) and they included the Bliss Sandstone as the lowest sequence. The sequences in this area range form 2 to 6 million years in duration. Within the third-order sequences, they recognized numerous higher order sequences at the scale of fourth- and fifth-order parasequences. These higher order sequences are the detailed depositional units and consist of meter-scale aggradational or progradational depositional cycles. This is the stratigraphic architectural scale that is used for flow-unit modeling in reservoir characterization (Kerans et al., 1994).

5

DEPOSITIONAL FACIES AND DEPOSITIONAL SYSTEMS Ellenburger Platform Systems Loucks and Anderson (1980, 1985) presented depositional models (Figures 10, 11) for the Ellenburger section in the Puckett Field in Pecos County of West Texas. Their data consisted of two cores that provide ~1700 ft of overlapping, continuous coverage of the section (Figure 12). They defined the lower Ellenburger section as being dominated by alluvial fan/coastal sabkha paracycles, the middle Ellenburger as subtidal paracycles, and the upper Ellenburger as supratidal/intertidal paracycles. Numerous forth- and fifthorder cycles occur within this Puckett Ellenburger section. They recognized many solution-collapsed zones that they attributed to exposure surfaces of different duration (Figure 12). Kerans (1990) completed the most detailed and complete regional Ellenburger depositional systems and facies analysis based on wireline log and core material. Much of the rest of this section is a summary of Kerans’ work. See Kerans (1990) for complete description and interpretation of facies)). He recognized six general lithofacies (Figure 7): (1) Litharenite: fan delta – marginal marine depositional system (2) Mixed siliciclastic-carbonate packstone/grainstone: lower tidal-flat depositional system (3) Ooid and peloid grainstone: high-energy restricted-shelf depositional system (4) Mottled mudstone: low-energy restricted-shelf depositional system (5) Laminated mudstone: upper tidal-flat depositional system (6) Gastropod-intraclast-peloid packstone/grainstone: open shallow-water depositional system Fan Delta – Marginal Marine Depositional System Description: Kerans (1990) noted that this system contains two prominent facies: cross-stratified litharenite and massive to cross-stratified pebbly sandstone to

6

conglomerate. Sedimentary structures include thick trough and tabular crossbeds, parallel current lamination, and graded and massive beds. Clastic grains are composed of granite and quartzite rock fragments, feldspar, and quartz. Interpretation: According to Kerans (1990) this unit was deposited as a fan delta – marginal marine depositional system. It is a basal retrogradational clastic deposit where the Ellenburger Group onlaps the Precambrian basement. Loucks and Anderson (1985) presented a similar interpretation of a fan-delta complex prograding into a shallow subtidal environment (Figure 11). Lower Tidal-Flat Depositional System Description: Kerans (1990) stated that the dominant facies are mixed siliciclasticpeloid packstone-grainstone, intraclastic breccia, stromatolitic boundstone containing silicified nodular anhydrite, ooid grainstone, and carbonate mudstone. These facies are mostly dolomitized. Kerans (1990) has noted that the siliciclastic content is related to the distribution of the sandstone below. Sedimentary structures include relict crossstratification, scour channels, stromatolites, flat cryptalgal laminites, and silica replaced evaporate nodules. Interpretation: According to Kerans (1990) this unit was deposited in a lower tidal-flat depositional system in close association with the fan delta depositional system (Figure 10). Upward in the section the carbonate tidal flats override the fan deltas. Kerans (1990) presented the idealized cycle within this system as an upward-shoaling succession. Tidal-flat complexes prograded across subtidal shoals and intervening lagoonal muds (Figure 10). The relic evaporate nodules indicate an arid sabkha climate (Loucks and Anderson, 1985; Kerans, 1990). Kerans (1990) pointed out that thin siliciclastic sand laminae in tidal-flat laminites represent eolian deposits, whereas thicker sand units represent periodic sheetflood deposits from adjacent alluvial fans. Loucks and Anderson (1985) also recognized the quartz sandstone units in the algal laminae.

7

High-Energy Restricted-Shelf Depositional System Description: Kerans (1990) noted that this system is characterized by ooid grainstone, ooid-peloid packstone-grainstone, laminated, massive, and mottled mudstone, and minor cyanobacterial boundstone. Also contains coarse-crystalline white chert and rare gastropod molds. Coarse dolomite fabric is common. Depositional structures include cross-stratification, intraclastic breccias, small stromatolites, cryptalgal mats, and silicified relict nodular anhydrite. Interpretation: According to Kerans (1990) this unit was deposited in a highenergy restricted-shelf depositional system. He stated that this system represents the period of maximum marine inundation during the Ellenburger transgression. He noted that extensive ooid shoals dominated the shelf and bioturbated mudstones formed in protected settings between shoals (Figure 11). Cryptalgal laminites and mudstones (tidal flats) with relic evaporate nodules may mark local shoaling cycles or more extensive upward shoaling events. Kerans (1990) noted that the lack of fauna suggests restricted circulation on the shelf produced by shoal-related restriction or by later destruction by dolomitization. Low-Energy Restricted Shelf Depositional System Description: Kerans (1990) described this widespread system as “remarkably homogeneous sequence of gray to dark-gray, fine- to medium-crystalline dolomite containing irregular mottling and lesser parallel-laminated mudstone and peloid wackestone.” He noted a sparse fauna of a few gastropods and nautiloids. The facies is highly dolomitized. Interpretation: According to Kerans (1990) this unit was deposited in a lowenergy restricted-shelf depositional system (Figure 11). The mottling is considered to be the result of bioturbation. This is a restricted shelf deposit ranging from subtidal mudstones to shoaling areas with tidal flats. Kerans (1990) noted that seaward, this system interfingers with the open-marine, shallow-water shelf depositional system. This fits the model of Ross (1976) (Figure 2).

8

Upper Tidal-Flat Depositional System Description: Kerans (1990) noted that the dominant facies in this system is smooth and parallel or irregular and crinkled laminated dolomite. Other facies include mottled mudstone, current-laminated dolostone, and beds of intraclastic breccia. Sedimentary structures include desiccation cracks, current laminations, nodular chert (relic evaporates?), and stromatolites. Interpretation: According to Kerans (1990) this unit was deposited in an upper tidal-flat depositional system. A common cycle is composed of a basal bioturbated mudstone, passing through current-laminated mudstone, and into cryptalgal laminated mudstone with desiccation structures and intraclastic breccias (Figure 11). Kerans (1990) noted that the mottled and current-laminate mudstones intercalated with the laminites, are low-energy shelf deposits and intercalated ooid-peloid grainstone beds are storm deposits transported from high-energy shoals offshore (Figure 11). Kerans (1990) suggested that the upper tidal-flat depositional system consisted of a broad tidal-flat environment situated landward of the lagoon-mud shoal complex. This is similar to the model presented by Loucks and Anderson (1985) (Figure 11). This depositional system occurs near the top of the Ellenburger succession. Open Shallow-Water Shelf Depositional System Description: Kerans (1990) noted that the rocks in this system are mainly limestone which is in contrast to much of the other sections of the Ellenburger interval. Facies include peloid and ooid grainstones, mollusc-peloidal packstones, intraclastic breccias, cryptalgal laminated mudstones, digitate stromatolitic boundstones, bioturbated mudstones, and thin quartzarenite beds. Again, in general contrast to the other depositional systems, this system has abundant fossils including sponges, trilobites, gastropods, bivalves, and cephalopods. Kerans (1990) described the grainstones and packstones as massive or displaying parallel current laminations. He noted abundant desiccation cracks in the laminites as well as fenestral fabric.

9

Interpretation: According to Kerans (1990) this unit was deposited in an open shallow-water shelf depositional system. He described the depositional setting as a complex mosaic of tidal-flat subenvironments, shallow-water subtidal carbonate sand bars, and locally thin stromatolite bioherms and biostromes (Figure 11). He interpreted the greater diversity of fauna, lack of evaporate evidence, and presence of high-energy grainstones and packstones to suggest a moderate-current energy environment with openmarine circulation. He speculated that this system may have occurred close to the shelf edge or slope break. Marathon Limestone Deeper Water System The Marathon Limestone is the time equivalent, deeper water slope facies of the Ellenburger shallow-water platform facies (Berry, 1960; Young, 1968; Ross et al., 1982; Kerans, 1990). In West Texas, it occurs in the Marathon Basin (Young, 1968) and on the western margin of the Diablo Platform (Lucia, 1968, 1969). Description: Kerans (1990) described the unit as containing graptolite-bearing shale, siliciclastic siltstone, lime grainstone and lime mudstone, and debris-flow megabreccia. Sedimentary structures consist of graded beds, horizontal laminations, sole marks, flute casts, and soft-sediment deformation structures (slump folds). Interpretation: According to Kerans (1990) this unit was deposited in a more basinal setting than the laterally equivalent Ellenburger depositional systems. He defined the setting as a distally steepened ramp. He recognized that the thin-bedded shale, siltstone, and lime grainstone-mudstone packages are Bouma turbidite sequences produced by turbidity currents on a deeper water slope. Both Young (1968) and Kerans (1990) interpreted the massively bedded megabreccias as deeper water debris-flow deposits. General Deposition History of the Ellenburger Group Kerans (1990) summarized the depositional history of the Ellenburger Group in four stages (Figure 13).

10

Stage 1: Marked by retrogradational deposition of fan delta – marginal marine depositional system continuous with the Early Cambrian transgression (Kerans, 1990). Kerans (1990) described the interfingering of the basal siliciclastics with the overlying the tidal-flat and shallow-water subtidal deposits of the lower tidal-flat depositional system. This transition represents initial transgression and associated retrogradational sedimentation. Kerans (1990) noted that this stage was followed by the regional progradation and aggradation of peritidal carbonate facies. This stage of deposition filled in existing paleotopography resulting in a low-relief shelf. Stage 2: Kerans (1990) documented rapid transgression and widespread aggradational deposition of the high-energy restricted-shelf depositional system across much of West and Central Texas during this stage. He noted that this transgression produced an extensive carbonate sand sheet over much of the platform. He interpreted a moderately hypersaline setting based on rare macrofauna, evidence of evaporites, and abundance of ooids. Stage 3: Kerans (1990) stated that the upward transition from the high-energy restricted shelf depositional systems to the low-energy restricted-shelf depositional systems is evidence of a second regression across the Ellenburger shelf. The progradation during this stage is marked by the transition of landward upper tidal flats, to more seaward low-energy restricted subtidal to intertidal facies, to furthest seaward openmarine, shallow-water shelf facies. Kerans (1990) recognized that the laminated mudstones of the upper tidal-flat depositional system represent the maximum regression across the Ellenburger inner shelf. Stage 4: Near the end of the Early Ordovician there was a worldwide eustatic lowstand. The timing of the lowstand is reported to be Whiterockian age (Sloss, 1963; Ham and Wilson, 1967). The length of exposure covered several million years. Throughout the United States, an extensive karst terrain formed on the Ellenburger platform carbonates (Kerans, 1988, 1989, 1990). During this long period of exposure, thick sections of cave development occurred resulting in extensive paleocave collapse breccias within the Ellenburger section (Lucia, 1971; Loucks and Anderson, 1980, 1985; Wilson, 1992; Kerans, 1988, 1989, 1990; Loucks 1999). The time equivalent, slopedeposited Marathon Limestone was not exposed during this sea-level drop (Kerans,

11

1990). The area appears to have had continuous deposition from the Lower Ordovician through the Middle Ordovician. GENERAL REGIONAL DIAGENESIS The diagenesis of the Ellenburger Group is extremely complex and the processes that produced the diagenesis covered millions of years (e.g., Folk, 1959; Lucia, 1971; Loucks and Anderson, 1985; Lee and Friedman, 1987; Kerans, 1988, 1989, 1990; Kupecz and Land, 1991; Amthor and Friedman, 1992; Loucks, 1999, 2003). Several studies have presented detailed diagenetic analysis of the Ellenburger (Kerans, 1990; Kupecz and Land, 1991; Amthor and Friedman, 1992). Several paragenetic charts are presented in Figures 14, 15. Three major diagenetic processes are important to discuss: (1) dolomitization, (2) karsting, and (3) tectonic fracturing. Other diagenetic features are present, but do not impact the appearance and reservoir quality of the Ellenburger as much as the three diagenetic events mentioned above. In the following discussion of these diagenetic processes, only an overview will be presented and the reader is referred to the literature on Ellenburger diagenesis for a complete and detailed discussion. Understanding Diagenesis in the Ellenburger Group As stated above, the diagenesis of the Ellenburger Group is complex. Detailed diagenesis can be worked out for any location, but trying to understand the complete diagenetic history for the entire Ellenburger carbonate section in West Texas may be beyond our reach because of relatively sparse subsurface data, long length of time (+/- 20 million years), thick stratigraphic section (possibly up to six third-order sequences), and the large area involved. It is important to keep in mind that carbonates generally undergo diagenesis very early in their history, especially if they are subjected to meteoric water. With the number of third-order sequences in the section and the time represented by each sequence (2 to 5 million years), extensive early and shallow diagenesis probably occurred but has been later masked by extreme dolomitization.

12

At the end of Early Ordovician time, a several million year hiatus occurred exposing the Ellenburger Group and subjecting it to meteoric karst processes. Several authors have demonstrated that the karst affected strata at least 300 to 1000 feet beneath the unconformity (e.g., Kerans 1988, 1989; Lucia, 1995; Loucks, 1999). With the occurrence of the Ouachita thrusting from Mississippian through Pennsylvanian, vast quantities of hydrothermal fluids moved though the available permeable pathways within the Ellenburger producing late stage diagenesis (e.g., Kupecz and Land, 1991). Following lithification, different parts of the Ellenburger Group were subjected to tectonic stresses and was fractured producing more late stage diagenesis that probably affected local areas (e.g., Loucks and Anderson, 1985; Kearns, 1990; Loucks, 2003). Loucks (2003) presented an overview of the origins of fractures in Ordovician strata and concluded that in order to understand the complex diagenesis in these strata, one must sort all the events into a well documented paragenetic sequence. This is the most reliable method to delineate timing of events and features. He was able to demonstrate that karsting and paleocave collapse breccias and related fractures and some tectonic fractures occurred before the hydrothermal events that produced saddle dolomite. This was accomplished by establishing well documented paragenetic relationships. Dolomitization Of the several authors (Kerans, 1990; Kupecz and Land, 1991; Amthor and Friedman, 1992) that have attempted to understand the regional dolomite history, Kupecz and Land (1991) appear to have made the most progress. This section will mainly address the finding of Kupecz and Land (1991), but still include observations and conclusions from the other authors. Kupecz and Land’s (1991) paragenetic sequence is presented in Figure 14. Their study covered a large area of West Texas as well as the Llano Uplift area in Central Texas. They used both cores and outcrop as a data source and combined petrography with carbon, oxygen, and strontium isotopes. They recognized five general stages of dolomitization (Figure 14). The generations of dolomite were separated into early-stage dolomitization which predated the Sauk unconformity and late-stage dolomitization that post dated the Sauk

13

unconformity. They attribute 90 percent of the dolomite as early stage and 10 percent of the dolomite as late stage. Kupecz and Land (1991) dolomite types: (1) Stage 1 pre-karstification early-stage dolomite (Dolomite E1) a. Description: Crystal size ranges from 5 to 700 μ but varies by facies. In the cryptalgal laminites it comprises crystal sizes ranging from 5 to 100 μ. These euhedral crystals have planar interfaces. In millimeter-laminated facies the crystal size ranges from 5 to 70 μ and in the bioturbated mudstones the crystal size ranges from 5 to 700 μ. Kupecz and Land (1991) thought that some of the coarser crystals were a product of later recrystallization. b. Interpretation: Kupecz and Land (1991) documented that this dolomite replaced lime mud or mudstone and that the dolomite predated karstification because it is found in nonkarsted rock as well as in clasts created by karsting. Therefore, it must have formed before karsting to be able to be brecciated. Probable source of Mg for dolomitization is sea water (Kupecz and Land, 1991). (2) Stage 2 post-karstification late-stage dolomite (Dolomite L1) a. Description: This replacement dolomite consists of coarse-crystalline euhedral rhombs with crystal size ranging from 200 to 2000 μ. Its homogeneous cathodoluminescence and homogeneous backscattered imaging suggest that this dolomite type has undergone recrystallization (Kupecz and Land, 1991). This stage of dolomitization is a regional event and is related to hydrothermal fluids. b. Interpretation: Late-stage origin is based on coarse-crystal size (Kupecz and Land, 1991). (3) Stage 3 post-karstification late-stage dolomite (Dolomite L2)

14

a. Description: Crystals have planar interfaces and size ranges from 100 to 3500 μ and has subhedral to anhedral crystal shapes. Extinction ranges from straight to undulose. b. Interpretation: It is a replacement type of dolomite (Kupecz and Land, 1991). Late origin is based on relationship to a later stage chert and its replacement of early-stage dolomite E1. Much of the grainstone facies is replaced by this stage of dolomitization. This stage of dolomitization is related to hydrothermal fluids. Probable source of Mg for dolomitization is dissolution of previous precipitated dolomite (Kupecz and Land, 1991). (4) Stage 4 post-karstification late-stage dolomite (Dolomite C1) a. Description: Crystals are subhedral with undulose extinction (saddle/baroque dolomite) and crystal size ranges from 100 to 5000 μ. b. Interpretation: Pore-filling cement (Kupecz and Land, 1991). Paragenetic sequence is established by the fact that Dolomite C1 post dates Dolomite L2 and was corroded before Dolomite Cement 2 was precipitated. Probable source of Mg for dolomitization is dissolution of previous precipitated dolomite (Kupecz and Land, 1991). This stage of dolomitization is a regional event and is related to hydrothermal fluids. (5) Stage 5 post-karstification late-stage dolomite (Dolomite C2) a. Description: Subhedral white crystals with moderate to strong undulose extinction (saddle/baroque dolomite) and crystal size ranges from 100 to 7500 μ. Contain abundant fluid inclusions. b. Interpretation: Pore-filling cement (Kupecz and Land, 1991). Occurred after corrosion of dolomite C1. Probable source of Mg for dolomitization is dissolution of previous precipitated dolomite (Kupecz and Land, 1991). This stage of dolomitization is related to hydrothermal fluids. Kupecz and Land (1991) have provided the only integrated analysis of fluid-flow pathways and sources of Mg for the different dolomitizing events. The early-stage prekarstification dolomite is associated with the muddier rocks and the source of Mg was probably sea water. Kerans (1990) similarly attributed these finer crystalline dolomites to

15

penecontemporaneous replacement of mud in tidal flats and to regionally extensive reflux processes during deposition. The late-stage post-karstification dolomites are attributed by Kupecz and Land (1991) to warm, reactive fluids, which were expelled from basinal shales during the Ouachita Orogeny. The fluids are thought to have been corrosive as evidenced by corroded dolomite rhombs (Kupecz and Land, 1991). This corrosion provided the Mg necessary for dolomitization. The warm, overpressured fluids were episodically released and migrated hundreds of miles from the fold belt toward New Mexico (Figure 15). These fluids migrated through high-permeability aquifers of the Bliss sandstone, basal subarkose facies of the Ellenburger, as wall as grainstone facies and paleocave breccia zones. Figure 15 from Kupecz and Land (1991) shows the regional isotopic composition of late-stage dolomite L2. The pattern of lighter to heavier delta-O18 away from the fold belt to the south, suggests movement and cooling of fluids to the northwest. Kupecz and Land (1991) regional dolomitization model is displayed in Figure 15. Figure 16 show the tectonic setting that produced the hydrothermal fluids. Kerans (1990) defined three major styles of dolomitization: (1) Very fine crystalline dolomite that he considered as a replacement product penecontemporaneous with deposition in a tidal-flat setting. (2) Fine- to medium-crystalline dolomite appearing in all facies and contributed to regionally extensive reflux processes during Ellenburger deposition. (3) Coarse-crystalline replacement mosaic dolomite and saddle (baroque) dolomite associated with burial. Kerans first two types of dolomite are probably equivalent to Kupecz and Land’s (1991) early-stage Dolomite E1. His coarse-crystalline replacement mosaic dolomite and saddle dolomite are equivalent to Kupecz and Land’s (1991) late-stage dolomites. Amthor and Friedman (1992) also recognized early- to late-stage dolomitization of the Ellenburger Group (Figures 17, 18). Similar to Kerans (1990) and Kupecz and Land (1991), Amthor and Friedman (1992) described early-stage, low-temperature, finecrystalline dolomites associated with lime muds where the Mg was supplied by diffusion from overlying seawater. Amthor and Friedman (1992) also describe medium- to coarse-

16

crystalline dolomite that replaces grains and matrix in the depth range of 1500 to 6000 ft. These dolomites are post-karstification and are probably replacement Dolomite L1 and L2 of Kupecz and Land (1991). Amthor and Friedman’s (1992) last stage of dolomite is assigned a deep-burial origin (>6000 ft) and consists of coarse-crystalline saddle dolomite. Its occurrence as both pore filling and replacive. This is Dolomite L2, C1, and C2 of Kupecz and Land (1991). Amthor and Friedman (1992) also noticed extensive corrosion of previously precipitated dolomite and they invoked a similar fluid-flow model as Kupecz and Land (1991) from which the fluids were associated with the Ouachita Orogeny. Overall, much of the Ellenburger is dolomitized. Dolomitization favors preserving open fractures and pores because it is mechanically and chemically more stable than limestone. Pores within dolomites are commonly preserved to deeper burial depths and higher temperatures than pores in limestone. Also, limestone breccia clasts tend to undergo extensive pressure solution at their boundaries and lose all interclast pores (Loucks and Handford, 1992), whereas dolomite breccia clasts are more chemically and mechanically stable with burial. Karsting Karsting is a complex and large-scale diagenetic event that strongly affected the Ellenburger Group. The process may affect only the surface of a carbonate terrain forming terra rossa or extensively dissolve the carbonate surface creating karst towers (Figure 19). It can also produce extensive subsurface dissolution in the form of dolines, caves, etc (Figure 19). The next several paragraphs are meant to provide background information on karst systems that are seen in the Ellenburger Group. Review of caves and paleocaves Loucks (1999) provided a review of paleocave carbonate reservoirs. He stressed that to understand the features of paleocave systems, an understanding of how paleocave systems form is necessary. The best approach to this understanding is to review how

17

modern cave systems form at the surface and evolve into coalesced, collapsed-paleocave systems in the subsurface. Loucks (1999) described this evolutionary process, and the review presented here is mainly from that investigation. To describe the features or elements of both modern and ancient cave systems, Loucks (1999) proposed a ternary classification of breccias and clastic deposits in cave systems based on relationships between crackle breccia, chaotic breccia, and cavesediment fill (Figure 20). Crackle breccias have thin fractures separating breccia clasts. Individual clasts can be fitted back together. Mosaic breccias are similar to crackle breccias, but displacement between clasts is greater and some clast rotation is evident. Chaotic breccias are characterized by extensive rotation and displacement of clasts. The clasts can be derived from multiple horizons, producing polymictic breccias. Chaotic breccias grade from matrix-free, clast-supported breccias to matrix-supported breccias. Loucks (1999) also showed that paleocave systems have complex histories of formation (Figure 21). They are products of near-surface cave development, including dissolutional excavation of passages, breakdown of passages, and sedimentation in cave passages. This is followed by later-burial cave collapse, compaction, and coalescence. Phreatic or vadose-zone dissolution create cave passages (Figures 19, 21). Passages are excavated where surface recharge is concentrated by preexisting pore systems, such as bedding planes or fractures (Palmer, 1991) that extend continuously between groundwater input, such as sinkholes, and groundwater output, such as springs (Ford, 1988). Cave ceilings and walls are under stress from the weight of overlying strata. A tension dome, a zone of maximum shear stress, is induced by the presence of a cavity (White, 1988). Stress is relieved by collapse of the rock mass within the stress zone. The collapse of the ceiling and wall rock commonly starts in the vadose zone. In the phreatic zone, water supplies 40% of the ceiling support through buoyancy (White and White, 1969). The removal of this support in the vadose zone weakens the ceiling and can result in its collapse. Major products of collapsed ceiling and walls are chaotic breakdown breccia on the floor of the cave passage (Figures 19, 21). In addition, the stress release around cave passages produces crackle breccias in the cave-ceiling and cave-wall host rocks (Figures 19, 21).

18

Near-surface dissolutional excavation and cave sedimentation terminate as cavebearing strata are buried into the subsurface. Extensive mechanical compacting begins, resulting in collapse of remaining passages and further brecciation of blocks and slabs (Figure 21). Multiple stages of collapse occur over a broad depth range, Foot-scale bit drops (cavernous pores) are not uncommon at depths of 6000 to 7000 ft (Loucks, 1999). The areal cross-sectional extent of brecciation and fracturing after burial and collapse is greater than that of the original passage (Figure 21). Collapsed, but relatively intact strata over the collapsed chamber are fractured and form burial cave-roof crackle and mosaic breccias with loosely to tightly fitted clasts (Figure 21). Sag feature and faults (suprastratal deformation) can occur over collapsed passages (Figure 21) (Lucia, 1971, 1995, 1996; Kerans, 1988, 1989, 1990; Hardage et al., 1996; Loucks, 1999, 2003). The development of a large collapsed paleocave reservoir is the result of several stages of development (Figure 22). The more extensive coalesced, collapsed-paleocave system originated at composite unconformities where several cave system may overprint themselves during several million years of exposure to karst process (Figure 22) (Esteban 1991; Lucia, 1995; Loucks, 1999). As the multiple-episode cave system subsides into the deeper subsurface, wall and ceiling rock adjoining open passages collapse and form breccias that radiate out from the passage and may intersect with fractures from other collapsed passages and older breccias within the system. This process forms coalesced, collapsed paleocave systems and associated reservoirs that are hundreds to thousands of feet across, and thousands of feet long, and tens to hundreds of feet thick. Internal spatial complexity is high, resulting from the collapse and coalescing of numerous passages and cave-wall and cave-ceiling strata. These breccias and fractures are commonly the major reservoirs in the Ellenburger Group. The reader is referred to Kerans (1988, 1989, 1990), Loucks and Handford (1992), Hammes (1996), and Loucks (1999, 2001, 2003) for discussions about paleocave systems in the Ellenburger Group. Loucks and Mescher (2001) have developed a classification of paleocave facies (Figure 23, Table 1). Six basic cave facies are recognized in a paleocave system and are classified by rock textures, fabrics, and structures: (1) undisturbed strata (undisturbed host rock), (2) disturbed strata (disturbed host rock), (3) highly disturbed strata (collapsed roof and

19

wall rock), (4) coarse chaotic breccia (collapsed-breccia cavern fill), (5) fine chaotic breccia (transported-breccia cavern fill), and (6) sediment fill (cave-sediment cavern fill). Each of paleocave facies can be distinct and adjoin sharply with adjacent facies, or they may show gradation into adjacent facies within the coalesced collapsed-paleocave system. Pore networks associated with paleocave reservoirs can consist of cavernous pores, interclast pores, crackle and mosaic breccia fractures, tectonic fractures, and less commonly, matrix pores. The paleocave facies classification, in conjunction with burial history data, can be used to describe the complex geology expressed in coalesced collapsed-paleocave systems and can be used to understand and predict pore-type distribution and magnitude of reservoir quality. Ellenburger Karsting In the Ellenburger Group, extensive cave systems formed at a composite unconformity (Sauk unconformity) that lasted several million years to several tens of million years. Many authors have recognized this karsting and associated features in the Ellenburger Group. Barnes et al. (1959) recognized solution collapse in the Ellenburger Group and stated that “… a matrix composed of material foreign to the formation indicates breccia formed by solution and collapse probably related to an erosional unconformity.” Lucia (1971) was the first to promote that the extensive brecciation seen in the El Paso Group (equivalent to Ellenburger Group) was associated with karst dissolution and were not the result of tectonic brecciation. Loucks and Anderson (1980, 1985) in the Puckett Field in Pecos County also realized that many of the breccias in the Ellenburger Group were related to solution collapse (Figure 11). They associated them with exposed diagenetic terrains. Kerans (1988, 1989, 1990) strongly established karsting and cave development in the Ellenburger Group. He proposed paleocave models (Figure 24) that were immediately accepted and applied. Lucia’s (1971, 1995, 1996) work in the El Paso Group in the Franklin Mountains of far West Texas presented an excellent outcrop analog for coalesced, collapsed paleocave systems. He has mapped a large paleocave system that was developed in the upper 1000 ft of the El Paso Group during a 33 million year time gap (Figures 25, 26).

20

Large fracture systems and collapse breccia zones of 1000 ft thick and 1500 ft wide, and over a mile long mark the collapsed paleocave system. Lucia (1995) noted that the cavernous porosity could have been as high as 30% before infilling with cave-sediment fill and cement. Within the southern Franklin Mountains, Lucia (1995) described the Great McKelligon Sag in McKelligon Canyon along the eastern face (Figure 27). The sag is ~1500 ft wide and ~150 ft deep. The sag formed by the collapse of paleocaves in the El Paso section after the Montoya and Fusselman units were deposited, buried, and lithified. This is an important feature representing the complete picture of a coalesced, collapsed paleocave system (Loucks, 1999). Loucks (2003) has stressed that the collapse of a coalesced paleocave system not only affects the karsted unit, but also strongly affects the units above (Figure 22). Loucks (2003) has called the deformation of the younger lithified units “suprastratal deformation.” Besides the example of suprastratal deformation shown in the Great McKelligon Sag, examples of suprastratal deformation from seismic in the Ellenburger Group can be seen in Hardage et al. (1996) and Loucks (1999, 2003) and from wireline-log cross sections by Kerans (1999) (his Figure 25). Kerans (1988, 1989, and 1990) presents an excellent overview of paleokarst in the Ellenburger Group. His paleocave models (Figure 24) were the first to define paleocave floor, paleocave sediment fill, and paleocave roof. Figures 28-32 presents several cores and associated core slabs from collapsed paleocave systems. These core descriptions emphasize Kerans’ paleocave model. The paleocave terrigenous-bearing sediment fill is strikingly apparent from the gamma-ray, spontaneous potential, and resistivity logs (Kerans, 1988, 1989, 1990). These paleocave fabrics can be recognized on electrical imaging tools (Hammes, 1997). Kerans (1988, 1989) discussed several breccia types: (1) a laterally persistent breccia association formed in the upper phreatic zone (water-table) karst and (2) a laterally restricted breccia association formed by deep phreatic dissolution and collapse. Kerans (1990) in his sequence of diagenetic events (his Figure 37) noted that the Ellenburger section was subjected to karsting during several periods of time. The main karst event was at the Early Ordovician Sauk-Tippecanoe Supersequence boundary. In local areas it was again karsted several more times. In the Llano area of Texas, Kupecz

21

and Land (1991) showed that the Ellenburger stayed near the surface until deep burial during later Pennsylvanian subsidence (Figure 33A). Loucks et al. (2000) recognized conodonts in cave-sediment fill from paleocaves in the Llano that would indicate exposure during the Middle to late Ordovician, Late Devonian, and Earliest Mississippian times. Also, they established other strong periods of karsting during the Pennsylvanian, Cretaceous, and Tertiary times. Combs et al. (2003) noted a second period of karsting in the Ellenburger interval in the Barnhart Field in Regan County where the Wolfcamp clastic overlies the Ellenburger Group. Their burial history diagram (Figure 33B) displays the two periods of karsting. Tectonic Fracturing In the past there has been controversy on the origin of many of the breccias and fractures in the Ellenburger Group. Some workers (e.g., Ijirighoi and Schreiber, 1986) wanted to assign most, if not all fractures, and breccias to a tectonic origin. They believed that faulting could produce these widespread breccias. The extensive size and shapes of most of the Ellenburger breccias and the inclusion of cave-sediment fill, speleothems, and younger conodonts, preclude a simple tectonic origin (Loucks, 2003). However, several authors have noted tectonic fractures in the Ellenburger section (e.g., Loucks and Anderson, 1985; Kearns, 1989; Holtz and Kerans, 1992; Combs et al., 2003). Each of the authors recognized that the tectonic fractures cut across lithified breccias. Kerans (1989) noted the fractures cutting late saddle dolomite. Kerans (1989) suggested that the Pennsylvanian foreland deformation that affected much of West and Central Texas, as described by Budnik (1987), probably produced many tectonic related fractures. However, as pointed out by many authors, paleocave collapse also can produce fractures that are not associated with tectonic events. Kerans (1990), Lucia (1996), and Loucks (1999, 2003) showed that suprastratal deformation above collapsing paleocave systems can create sags, faults, and numerous fractures (Figure 22). Kerans (1989) pointed out several distinct ways to separate karst-related fractures from tectonic-related fractures:

22

(1) Tectonic fractures are commonly the youngest fractures in the core and generally cross cut karst-related fractures. (2) Tectonic fractures post-date saddle dolomite. In the Llano area, however, saddle dolomite fills in a well-developed Pennsylvanian fracture set (Loucks, 2003). (3) Karst-related fractures are generally near the top of the Ellenburger section, whereas tectonic-related fractures can occur throughout the Ellenburger section. Loucks and Mescher (1998) presented additional criteria for separating karst-related fractures from tectonic-related fractures: (1) Tectonic fractures generally show a strong relationship to regional stress patterns and have well defined oriented sets of fractures, whereas karst-related fractures respond to near-field stresses and fracture orientation is more random than tectonic fractures. (2) Regional tectonic fractures are usually spaced at greater than the inch scale and commonly at the foot or larger scale. Karst-related fractures (crackle breccia fractures can be very closely spaced only a fraction of an inch apart. (3) Breccias associated with tectonic derived faults commonly from a narrow band around the fault only a few feet wide but may be tens of feet wide. Karst-related breccias can be thousands of feet wide and hundreds of feet thick, contain a large range of clast sizes, and show hydrodynamic sedimentary structures. (4) Tectonic faults are linear or curved in map view. Karst-related faults are linear, curved, or cylindrical (Hardage et al., 1996; Loucks, 1999) in map view. Tectonic- and karst-related fractures are both present in the Ellenburger section. Detailed analysis of the fractures usually can define their origin.

23

RESERVOIR CHARACTERISTICS Pore Types Pore networks in the Ellenburger are especially complex because of the amount of dolomitization, brecciation and fracturing associated with karsting, and regional tectonic deformation. The pore networks can consist of any combination of the following pore types depending on depth of burial (Loucks, 1999): (1) matrix, (2) cavernous, (3) interclast, (4) crackle/mosaic breccia fractures, or (5) tectonic-related fractures. Loucks (1999) presented an idealized plot of how karst-related Ellenburger pore networks probably change with depth (Figure 34). Relative abundance of pore types and relative depth of burial are estimates based on review of near-surface modern cave systems and buried paleocave systems (see Table 2 in Loucks, 1999). Large voids may be preserved down to 8000 to 9000 feet of burial, but eventually collapse forming smaller interclast pores and fractures associated with crackle and mosaic breccias. Coarseinterclast pores between large clasts are reduced by rotation of clasts to more stable positions and by rebrecciation of clasts to smaller fragments (Figure 21). As passages and large interclast pores in the cave system collapse, fine-interclast pores first increases and then decreases, whereas fracture pore types become more abundant. Cave-sediment fill is commonly cemented tight during burial diagenesis in the Ellenburger carbonates, especially if it is terrigenous sediment or has a terrigenous sediment component. In the Ellenburger carbonates, matrix porosity is generally low (